Home > Smartphone processors comparison > Snapdragon 8 Gen 2 vs Tensor G3 – what's better?

Snapdragon 8 Gen 2 vs Tensor G3

Snapdragon 8 Gen 2
VS
Tensor G3
Snapdragon 8 Gen 2
Tensor G3

We compared the 8-core Qualcomm Snapdragon 8 Gen 2 (Adreno 740) with the newer 9-core Google Tensor G3 (Mali-G715 MP7) SoC. Here you will find the pros and cons of each chip, technical specs, and comprehensive tests in benchmarks, like AnTuTu and Geekbench.

  1. Review
  2. Differences
  3. Benchmarks
  4. AnTuTu v11
  5. GeekBench 6
  6. Gaming
  7. Specs
  8. Comments (13)

Review

General comparison of performance, power consumption, and other indicators
CPU Performance
Single and multi-core processor tests
Gaming Performance
GPU performance in games and OpenCL/Vulkan
Battery life
Efficiency of battery consumption
NanoReview Score
Overall chip score

Key Differences

Main differences and advantages of each chip
Pros of Qualcomm Snapdragon 8 Gen 2
  • Performs 31% better in floating-point computations
  • Shows better (up to 25%) AnTuTu 11 score – 1757K vs 1409K
  • 10% higher CPU clock speed (3200 vs 2910 MHz)
Pros of Google Tensor G3
  • Higher GPU frequency (~31%)
  • Has 1 more cores

Benchmarks

Performance tests in popular benchmarks
SoC:
Snapdragon 8 Gen 2
vs
Tensor G3

AnTuTu 11

The AnTuTu Benchmark measures CPU, GPU, RAM, and I/O performance in different scenarios
Snapdragon 8 Gen 2 +25%
1757044
Tensor G3
1409016
CPU 553922 482403
GPU 434880 384611
Memory 311861 218603
UX 456381 323399
Total score 1757044 1409016
Samples [238 tests] [33 tests]
Submit your AnTuTu result

GeekBench 6

The GeekBench test shows raw single-threaded and multithreaded CPU performance
Single-Core Score
Snapdragon 8 Gen 2 +16%
2004
Tensor G3
1722
Multi-Core Score
Snapdragon 8 Gen 2 +19%
5311
Tensor G3
4458
Asset compression 218.9 MB/sec 197.6 MB/sec
HTML 5 Browser 145.2 pages/sec 134.7 pages/sec
PDF Renderer 168.3 Mpixels/sec 148.4 Mpixels/sec
Image detection 133 images/sec 82.1 images/sec
HDR 167.2 Mpixels/sec 142.1 Mpixels/sec
Background blur 18.8 images/sec 13.2 images/sec
Photo processing 34.5 images/sec 48.4 images/sec
Ray tracing 5.78 Mpixels/sec 5.34 Mpixels/sec
Samples 132 tests 11 tests
Compute Score (GPU)
Snapdragon 8 Gen 2 +15%
9254
Tensor G3
8076
Background blur 12.3 images/sec 10.6 images/sec
Face detection 5.1 images/sec 6.85 images/sec
Horizon detection 0.53 Gpixels/sec 0.16 Gpixels/sec
Edge detection 0.56 Gpixels/sec 0.35 Gpixels/sec
Gaussian blur 0.36 Gpixels/sec 0.32 Gpixels/sec
Feature matching 0.12 Gpixels/sec 0.08 Gpixels/sec
Stereo matching 32.6 Gpixels/sec 11.9 Gpixels/sec
Particle physics 2177.4 FPS 1088 FPS
API OpenCL Vulkan
Samples 66 tests 9 tests

3DMark

A cross-platform benchmark that assesses graphics performance in Vulkan (Metal)
3DMark Steel Nomad Light
Score 1092 982
Stability 67% 74%
Graphics test 8 FPS 7 FPS
3DMark Solar Bay
Score 5312 -
Stability 67% -
Graphics test 20 FPS -
Section 1 23 FPS -
Section 2 20 FPS -
Section 3 17 FPS -
3DMark Wild Life Extreme
Score 3667 -
Stability 45% -
Graphics test 21 FPS -

PCMark 3.0

Benchmark for evaluating overall system performance in everyday workloads
PCMark 3.0
Web score 13957 -
Video editing score 7265 -
Data manipulation score 14272 -
Photo editing score 32179 -
Writing score 17828 -
Phones tested 9 models -

Gaming

Table of average FPS and graphics settings in mobile games
PUBG Mobile 90 FPS
[Ultra]		 -
Call of Duty: Mobile 60 FPS
[Ultra]		 -
Fortnite 56 FPS
[Ultra]		 -
Shadowgun Legends 118 FPS
[Ultra]		 -
World of Tanks Blitz 120 FPS
[Ultra]		 -
Genshin Impact 59 FPS
[Ultra]		 -
Mobile Legends: Bang Bang 59 FPS
[Ultra]		 -
Device ZTE Nubia Red Magic 8 Pro
1116 x 2480		 -
We provide average results. FPS may differ, depending on game version, OS and other factors.

Recent User Tests

The latest benchmark tests that have been submitted by users
Qualcomm Snapdragon 8 Gen 2
DateBenchmarkResult
πŸ“˜ 2026-09-23 (Ciciz)AnTuTu 122032599
πŸ“˜ 2026-09-23 (Ciciz)AnTuTu 122032599
πŸ“˜ 2026-09-23 (Horizon377)AnTuTu 121894819
πŸ“˜ 2026-09-22 (DiKevoLi'Phone)AnTuTu 121844320
πŸ“˜ 2026-09-21 (Mil)AnTuTu 121763249
Add View all (368)
Google Tensor G3
DateBenchmarkResult
πŸ“˜ 2026-08-28 (asfiii-malik)AnTuTu 111515235
πŸ“˜ 2026-08-28 (asfiii-baloch)AnTuTu 111515235
πŸ“˜ 2026-05-05 (s1mpick)AnTuTu 111338293
πŸ“˜ 2026-05-04 (reazz)AnTuTu 111356451
πŸ“˜ 2026-05-04 (Bilal Ahmed)AnTuTu 111367231
Add View all (48)

Specifications

Full list of technical specifications of Snapdragon 8 Gen 2 and Tensor G3

CPU

Architecture 1x 3.2 GHz – Cortex-X3
2x 2.8 GHz – Cortex-A715
2x 2.8 GHz – Cortex-A710
3x 2 GHz – Cortex-A510		 1x 2.91 GHz – Cortex-X3
4x 2.37 GHz – Cortex-A715
4x 1.7 GHz – Cortex-A510
Cores 8 9
Frequency 3200 MHz 2910 MHz
Instruction set ARMv9-A ARMv9-A
L2 cache 1 MB -
L3 cache 8 MB -
Process 4 nanometers 4 nanometers
Transistor count 16 billion -
TDP (Sustained Power Limit) 8.5 W 6 W
Manufacturing TSMC Samsung

Graphics

GPU name Adreno 740 Mali-G715 MP7
Architecture Adreno 700 Valhall 4th gen
GPU frequency 680 MHz 890 MHz
Pipelines 3 7
Shading units 512 128
Total shaders 1536 896
FLOPS 2089 Gigaflops 1594.8 Gigaflops
Vulkan version 1.3 1.3
OpenCL version 2.0 2.0
DirectX version 12.1 -

AI Accelerator

Neural processor (NPU) Hexagon Yes

Memory

Memory type LPDDR5X LPDDR5X
Memory frequency 4200 MHz 4266 MHz
Bus 4x 16 Bit 4x 16 Bit
Max bandwidth 67.2 Gb/s 68.2 Gb/s
Max size 24 GB 16 GB

Multimedia (ISP)

Storage type UFS 3.1, UFS 4.0 UFS 3.1, UFS 4.0
Max display resolution 3840 x 2160 3840 x 2160
Max camera resolution 1x 200MP 1x 200MP, 2x 32MP
Video capture 8K at 30FPS, 4K at 120FPS 8K at 30FPS, 4K at 120FPS
Video playback 8K at 60FPS, 4K at 120FPS 8K at 30FPS, 4K at 120FPS
Video codecs - H.264
- H.265
- AV1
- VP8
- VP9		 - H.264
- H.265
- AV1
- VP9
Audio codecs - AAC
- AIFF
- CAF
- MP3
- MP4
- WAV		 - AAC
- AIFF
- CAF
- MP3
- MP4
- WAV

Connectivity

Modem Snapdragon X70 Exynos 5300i
4G support LTE Cat. 24 LTE Cat. 24
5G support Yes Yes
Download speed Up to 10000 Mbps Up to 7300 Mbps
Upload speed Up to 3500 Mbps Up to 3700 Mbps
Wi-Fi 7 7
Bluetooth 5.3 5.3
Navigation GPS, GLONASS, Beidou, Galileo, QZSS, SBAS, NAVIC GPS, GLONASS, Beidou, Galileo, QZSS

Info

Announced November 2022 October 2023
Class Flagship Flagship
Model number SM8550-AB GS301 (S5P9865)
Official page Qualcomm Snapdragon 8 Gen 2 official site Google Tensor G3 official site
Successor - Qualcomm Snapdragon 8 Elite Gen 5 - Google Tensor G5
Further details
Notes on Snapdragon 8 Gen 2:
    - There's also the "Snapdragon 8 Gen 2 for Galaxy," which only differs by having a 160 MHz higher CPU clock and a 39 MHz higher GPU clock, resulting in 2208.8 GFLOPS for the GPU. Generally, you won't notice this difference in real-life usage, including games.

Cast your vote

So, which SoC would you choose?
1508 (76%)
476 (24%)
Total votes: 1984

Related Comparisons

1. Qualcomm Snapdragon 8 Gen 2 vs Qualcomm Snapdragon 8s Gen 3 (106)
2. Qualcomm Snapdragon 8 Gen 2 vs Qualcomm Snapdragon 8 Elite (Gen 4) (19)
3. Qualcomm Snapdragon 8 Gen 2 vs MediaTek Dimensity 8300 (36)
4. Qualcomm Snapdragon 8 Gen 2 vs Qualcomm Snapdragon 7 Plus Gen 3 (18)
5. Qualcomm Snapdragon 8 Gen 2 vs MediaTek Dimensity 8400 (207)
6. Google Tensor G3 vs Qualcomm Snapdragon 8 Gen 3 (3)
7. Google Tensor G3 vs Google Tensor G2 (6)
8. Google Tensor G3 vs Google Tensor
9. Google Tensor G3 vs Google Tensor G4 (18)
10. Google Tensor G3 vs Qualcomm Snapdragon 8 Gen 1 (7)
Compare other chipsets (190+)

Comments

Please give your opinion on the comparison of Tensor G3 and Snapdragon 8 Gen 2, or ask any questions
🌐 Register your profile and become part of NanoReview community!
Avatar
kem April 25, 2025 at 06:26 AM
Qualcomm trued samsung and it fks up their SOC 888 and 8 gen 1 now they go back to tsmc i think Tensor should do the same Samsung fabrication is like where they’re trying but still mess up shame to samsung selling low performing soc as same price to high performing qualcomm soc if they lower the price it will considerable but theyre selling experimental product at same price as flagship products thats mess up i think
+4 Reply
Avatar
Icuka August 12, 2024 at 08:51 PM
This is right here why I wont even consider the Pixel series, until they stop using Samsung's F tier SoC. I mean even Samsung only dares to use it in the EU where apparently ppl are clueless enough to buy it without any second thought.
+3 Reply
Avatar
Pixel user April 17, 2025 at 08:39 PM
It's not that people are clueless to buying a exynos device, it's that it's all they can buy, here in Europe we don't get a choice, it's exynos or nothing.
+3 Reply
Avatar
Mrsch July 1, 2024 at 08:39 AM
I'm baffled about the fact that "Video Boost" is done on Google's servers, rather than on-device. With the AI focus they're advertising on Tensor G3... So disappointed...
0 Reply
Avatar
Paul May 8, 2024 at 09:29 AM
It's been clear to me for a while now that smartphone performance between the flagships of major manufacturers is starting to be very similar. Differences relate to decisions made to favor certain aspects in the instruction set, cores, etc. Huge leaps ahead of competitors are unlikely unless someone makes another major leap forward, i.e., Apple's switch from Intel to ARM in the M series. I honestly believe it's good for customers to see manufacturers have more incremental improvements. For example, Google focusing on the Exynos Tensor G series is why they can now offer 7 years of support. Apple has done that with the A series. That allows developers the ability to innovate on the understanding of a solid foundation aligned to a future strategy of platform innovation. Providing developers with access to more open standards is good for all and I wish all manufacturers would support that. I'm looking at you, Apple and Qualcomm!
+7 Reply
Avatar
New comer January 20, 2024 at 08:07 AM
Google is a newcomer on Cpu. With the short time, they catch up pretty fast. Snapdragon has been in the market long.
+3 Reply
Avatar
Dios April 9, 2024 at 07:15 AM
No, they aren't. They are using Exynos SoCs as a foundation. This is precisely why it's as disappointing as Exynos has ever been.
+20 Reply
Avatar
King Miracle. August 6, 2024 at 02:35 PM
Dios, They're using Exynos to go into R and D. When they can do their own designs, they will move to TSMC. Even Qualcomm did that too.
0 Reply
Avatar
Lol November 9, 2023 at 08:48 PM
The Tensor G3 is laughable, and it becomes an even bigger joke when the 8th Gen 3 is released. It even puts Google's NPU and AI to shame.
+50 Reply
Avatar
Lolol November 14, 2023 at 02:45 PM
When I first saw these specs, I thought the same thing. That is, until I noticed that the 8 Gen 2 was consistently 40% faster in most cases - a typical percentage of improvement in speed for succeeding generation chips. It's quite evident that the Tensor G3 is designed to compete with the 8 Gen 1. Its scores are incredibly similar, and in some instances, slightly better than the 8 Gen 1.
+17 Reply
Avatar
lolx2 November 17, 2023 at 06:16 PM
Lolol, I reached the same conclusion. When Google releases its next-generation processor, I hope they anticipate competition with Snapdragon's upcoming next-generation model.
+10 Reply
Avatar
lolwa December 2, 2023 at 05:05 AM
lolx2, When I first saw these specs, I thought the same thing. That was, until I noticed that the Gen 2 8 was consistently 40% faster in most cases - a typical percentage of improvement for next generation chips. It's quite evident that the Tensor G3 is designed to compete with the Gen 1 8. Its scores are incredibly similar, and in some instances, slightly better than the Gen 1 8.
+22 Reply
Avatar
joseph April 19, 2024 at 05:23 PM
Actually when it comes to AI using android NNAPI delegate to run inference on Tensor G3's TPU is offer such a crazy speedup for on device AI inference. Meanwhile for Qualcomm, they have stopped giving access to Android NNAPI starting from their 8th gen 2 chipset and instead force developers like me to use their own QNN delegate that is closed source and can be accessed only by "verified companies". It's been almost 2 years and it's not even clear which companies can be considered verified. And even when we got access, the documentation, customer support and community is very lacking. Overall, from a standpoint of an AI researcher working on edge device AI, I much rather prefer and support google's platform over qualcomm's current platform state. However, I am of course open to change my mind in case there will be any updates soon.
+15 Reply
EnglishРусский